an appraisal of antiretroviral drugs in hemodialysis

Kidney International, Vol. 60 (2001), pp. 821–830

PERSPECTIVES IN RENAL MEDICINE

An appraisal of antiretroviral drugs in hemodialysis

HASSANE IZZEDINE, VINCENT LAUNAY-VACHER, ALAIN BAUMELOU, and GILBERT DERAY

Department of Nephrology, Pitie Salpetriere Hospital, Paris, France

An appraisal of antiretroviral drugs in hemodialysis. HEMODIALYSIS AND DRUG ELIMINATIONBackground. Acquired immunodeficiency syndrome (AIDS)-

Dialysis therapy is used in chronic uremia to removerelated kidney disorders concern 30% of those patients and cantoxic waste products that accumulate in patients withlead to end-stage renal disease (ESRD; 6 to 10%). Therefore,

the administration of antiretroviral drugs in human immunodefi- ESRD. However, this technique also removes drugs asciency virus (HIV) patients with nephropathy is not uncommon. well as waste. Therefore, it is necessary to determine

Methods. The influence of ESRD on the different phases of the removed fraction of drugs during the procedure bythe pharmacokinetic profile of drugs in general is examined in

hemodialysis clearance, extraction coefficient, and he-light of bioavailability, distribution, protein binding, metabo-modialysis elimination of the absorbed dose (FHD).lism, and elimination. Then, the pharmacokinetics of antiret-

roviral drugs in hemodialysis are detailed.Hemodialysis clearanceResults. From these data, dosing recommendations are given

for nucleoside reverse transcriptase inhibitors (NRTIs), non- Dialysis clearance is the rate of removal relative toNRTIs, and protease inhibitors (PIs).

the concentration in the blood entering the dialyzer. ItConclusion. Dosage adjustments are often necessary for pa-is calculated by the amount removed (mg/min) relativetients with renal insufficiency. These adaptations have to be

carefully performed to optimize drug exposure and reduce the to the rate of presentation (mg/mL).risk of side effects.

CHD (mL/min) � [(Ca � Cv) � Qb]/Ca

where CHD is the hemodialysis (HD) clearance, Ca is theRenal insufficiency is common in human immuno- concentration entering the dialyzer (mg/mL), Cv is the

deficiency virus (HIV)-infected patients (6 to 10%) [1, 2]. concentration leaving the dialyzer (mg/mL), and Qb isHIV-associated nephropathy (HIVAN) and other ac- the blood flow (mL/min).quired immunodeficiency syndrome (AIDS)-related kid-

Extraction coefficientney disorders are expected to become leading causesof end-stage renal disease (ESRD). Nucleoside reverse The extraction coefficient (E), also called the extrac-transcriptase inhibitors (NRTIs) used in HIV therapy tion ratio, is the percentage of a drug removed fromare primarily excreted by the kidney. Furthermore, they blood across the dialyzer. It is calculated by the rate ofare most often used in combination: two NRTIs plus one removal (mL/min) relative to the rate of presentationprotease inhibitor (PI) or two NRTIs plus one non-NRTI. (mL/min).Therefore, clinicians need to understand the pharmaco-

E (%) � CHD/Qbkinetic disposition of these compounds in patients withrenal impairment. where CHD is the hemodialysis clearance (removal; mL/

This article reviews the impact of hemodialysis on the min), and Qb is the blood flow rate (mL/min).pharmacokinetics of the antiretroviral drugs used in HIV Dialysis clearance and extraction coefficient measuredisease and discusses the dosage recommendations needed the ability of a dialysis system to remove a drug fromto achieve efficacy while avoiding toxicity in patients the blood. Nevertheless, they do not indicate how readilywith ESRD undergoing hemodialysis. the drug is removed from the body [3].

Drug administration in hemodialysis patientsKey words: dosing recommendations, AIDS, HIV, nucleoside reverse

To evaluate whether a drug is significantly removedtranscriptase inhibitors, protease inhibitors, end-stage renal disease,dialysis. by dialysis and then to know what supplemental dose

should be administered at the end of the session, oneReceived for publication March 29, 2000approach consists of replacing the amount lost in theand in revised form March 8, 2001

Accepted for publication March 15, 2001 dialysate during the treatment period. This amount canbe calculated from the amount in the body at the start 2001 by the International Society of Nephrology

821

Izzedine et al: Pharmacokinetics of antiretroviral drugs822

of dialysis: V*C0. The second existing approach would of the drug to the dialysis membrane. Therefore, drugbe more accurate and consists of restoring the concentra- removal by dialysis is insignificant for agents with largetion of the drug at the end of the dialysis session to the Vd values. Protein binding may help determine where avalue that would have occurred if the patient had not drug is distributed once in the body. The degree of pro-been dialyzed. The amount in body without dialysis can tein binding of a drug also has a significant impact onbe estimated with V*C0*exp(�k*t), with k being the its pharmacological effects. Patients with renal diseaseelimination rate of the drug without dialysis. The amount exhibit less protein binding of various drugs than subjectsin the body after the dialysis session can be estimated with normal renal function. Qualitative [6] and quantita-with V*C0*exp(�kD*T), with kD being the elimination tive [7] changes in the albumin level of uremic subjectsrate of the drug during the session. Then, the supplemen- explain these observations.tary dose can be estimated with V*C0*[exp(�k*T) � Elimination. Elimination is the irreversible loss ofexp(�kD*T)]. drug from the site of measurement (blood, plasma, etc.).

Very often, these data are not available to the clinician. It occurs by two processes: excretion and metabolism [8].However, the influence of dialysis may be derived from In renal impairment, a drug usually cleared by renaltotal body clearance (Ctot) and hemodialysis clearance mechanisms will have a longer half-life (T1⁄2) and an in-(CHD). Hemodialysis elimination of the absorbed dose creased Vd.(FHD) is assumed to be clinically relevant if more than Metabolism. Metabolism is the conversion of one chemi-25% of the overall elimination of a drug in a hemodia- cal species to another [8]. Generally, drugs that undergolyzed patient is by this route [4]: FHD � CHD/Ctot. Since hepatic metabolism by microsomal oxidation have simi-this is a quite recent concept, FHD was not reported in lar elimination profiles in patients with normal and im-these studies, so we calculated it from data on CHD and paired renal function [9].Ctot found in the text of these studies. Ctot consists of Excretion. Excretion is the irreversible loss of a chem-the sum of hemodialysis clearance (CHD) and total body ically unchanged drug [8]. In dialysis patients, renal ex-clearance derived from the pharmacokinetic analysis on cretion and drug removal by the kidney are replaceda non-hemodialysis day (CtotNHDD). As a result, FHD �

by dialysis. Drugs normally eliminated via glomerularCHD/(CHD � CtotNHDD). Most of the time, those data were

filtration will be at least partially dialyzable. Other drugs,available. To assess FHD, Ctot and CHD must be calculatednormally eliminated via tubular secretion, may or mayon a nondialysis day and during dialysis, respectively. FHD not be dialyzable, depending on their molecular weight.is reported as unknown when those data were lacking.

It must be emphasized that hemodialysis clearance Drug properties affecting dialyzabilityand the extraction ratio (without FHD) do not allow any

Molecular weight. A drug’s molecular weight is a reli-extrapolation to assess which supplemental dose is war-able predictor of its dialyzability [10]. If a drug cannotranted.pass through a dialysis membrane pore because of itsphysical size, it is not cleared by the dialyzer [11]. Diffu-PHARMACOKINETICS IN HEMODIALYSISsive clearance through conventional dialysis membranesIn patients with ESRD, not only elimination, but allis negligible for molecules larger than 1000 D. Very fewphases of drug pharmacokinetics may be altered.commonly used drugs are this large. Utilizing conven-tional low-flux cellulosic membranes, removal of solutesGeneral considerationsgreater than 1000 D generally requires ultrafiltration.Bioavailability. Gastric pH is frequently modified inDepending on the membrane, even during ultrafiltration,uremic patients. Dialysis patients have an increased sali-molecules greater than 2000 D are partially reflected byvary urea concentration that is converted into ammoniathe membrane [12].by gastric ureases. Drugs requiring an acid environment

Protein binding and Vd. For drugs characterized byfor optimal absorption may thus exhibit a decreased ab-high protein binding, only a small proportion of the drugsorption ratio in uremic settings.is available for removal by dialysis. Heparin administra-Distribution. Uremia alters the apparent volume oftion during hemodialysis stimulates lipoprotein lipase,distribution (Vd) by various mechanisms. Endogenouswhich in turn raises the levels of free fatty acids [13].inhibitors of binding are retained in renal failure andFree fatty acids may displace or enhance the binding oflead to a larger Vd [5]. Drugs that are minimally bounddrugs. Furthermore, when a drug has a large Vd, it diffusesto plasma proteins have the same Vd in uremia as inwidely in the tissues, and its availability in the circulationsubjects with normal renal function [6]. When systemicand for the dialyzer is minimal [14]. A comparison ofacidosis is associated with uremia, it allows an increaseddrugs with various Vd sizes and same molecular weightsdrug penetration into the central nervous system (CNS).and protein-bound fractions shows that drugs with a VdThe larger the Vd, the smaller the proportion of drug

remains in circulation and hence the smaller the delivery below 1 L/kg are more readily dialyzed and those with

Izzedine et al: Pharmacokinetics of antiretroviral drugs 823

a Vd from 1 to 2 L/kg have marginal dialysis clearances; tively) [16, 17]. This is consistent with the physicochemi-cal and kinetic properties of the drug (molecular weightfor those with a Vd greater than 2 L/kg, substantial drug

removal is unlikely. Despite rapid extracellular solute �1000 D, and low protein-bound fraction and volumeof distribution). ZDV is not significantly removed byclearance with short-time high-flux or high-efficiency di-

alysis, intracellular equilibration with extracellular fluid hemodialysis (FHD 5 to 8%) in contrast with GZDV, thehemodialysis clearance of which is three to four timescan be slow, especially with larger molecules. This is prob-

ably related to the lipid solubility of the drug and its the residual clearance in nondialyzed uremic patients.However, in the absence of any data about GZDV activ-tissue compartmentalization. A postdialysis extracellular

rebound usually occurs (10 to 25% variations), but with ity and/or toxicity, it is recommended that the doseshould be administered after the dialysis session to avoidlow influence on intracellular drug concentrations (1 to

2% variations). Higher filtration rates aggravate this re- potential clinically significant removal of the metabolite.Didanosine. Didanosine (ddI) has a molecular weightbound phenomenon.

Drug-red blood cell binding. Drugs that diffuse in red of 236.23 D, a fraction bound to plasma protein of lessthan 5%, and a volume of distribution of 0.8 L/kg. Fiftyblood cells (RBCs) and have a partition coefficient ex-

ceeding one unit may have decreased clearances due to to 55% of each ddI dose is excreted unchanged in theurine. The nonexcreted fraction is metabolized by thehemoconcentration at the end of dialysis.

Miscellaneous drug properties. Water solubility and liver into hypoxanthine, uric acid, and dideoxyadenosinetriphosphate [17]. Certain changes in the pharmacokinet-electric charge of a drug may affect its dialyzability. It

has been suggested that charged drugs are less dialyzable ics of ddI occur in renal impairment. The area under theconcentration-time curve (AUC) is fourfold to fivefoldthan uncharged. However, the data on real impact of

these properties on drug dialyzability are scanty. greater than in subjects with normal renal function. Dos-age adjustment is therefore necessary for patients withrenal insufficiency (Table 5). Hemodialysis clearance of

CLINICAL APPLICATIONS INddI is relatively high (85 or 210 mL/min using a cupro-

HEMODIALYZED PATIENTSphan or a polysulfone membrane, respectively), with an

Figure 1 illustrates the molecular structure of the anti- extraction ratio of 40 to 60% during a four-hour dialysisviral drugs. The following dosing recommendations of session and a FHD of 40%. In hemodialyzed patients, thethese drugs in hemodialysis patients proceed from the daily dose should be administered after the dialysis sessionanalysis of the existing pharmacokinetic studies, the to minimize drug loss [18]. This is consistent with themethodologies of which are summarized in Table 1. Re- physicochemical and kinetic properties of the drug (mo-nal abnormalities induced by antiretroviral drugs are lecular weight �1000 D, and low protein-bound fractionsummarized in Table 2. and volume of distribution).

Zalcitabine. Zalcitabine (ddC) has a molecular weightNucleoside reverse transcriptase inhibitors of 211.22 D, a fraction bound to plasma protein less than

Protein binding of NRTIs is low. The free fraction 4% and a volume of distribution of 0.5 L/kg. Seventy toof drug is unchanged. Drug metabolism is mediated by 75% of the dose of ddC is excreted unchanged in thevarious cellular kinases and endogenous nucleotide tri- urine. Tubular secretion contributes significantly to uri-phosphate pools [15]. Ten to 60% of NRTIs is usually nary elimination. The half-life of the parent drug is five-excreted unchanged by the kidney. Thus, the unchanged fold higher than normal in patients with renal failure [19].form and metabolites of accumulate in patients with re- The pharmacokinetics of ddC have not been studied innal insufficiency and may cause adverse effects. The dialysis patients. However, its dosage should be the samelower a drug’s molecular weight, the more significant its as for patients with creatinine clearance below 10 mL/removal by dialysis (Tables 3 and 4). min, and it should be administered after the hemodialysis

Zidovudine. Zidovudine (ZDV) has a molecular weight session to avoid any elimination in the dialysate (Table 5)of 267.24 D, a fraction bound to plasma protein less than [20]. Indeed, the physicochemical and kinetic data (molecu-38% and a volume of distribution of 1.5 L/kg. In patients lar weight �1000 D, low protein-bound fraction and vol-with normal renal function, 15 to 20% of ZDV is excreted ume of distribution) suggest that ddC may be dialyzable.unchanged by the kidney. Sixty to 75% of the drug recov- Stavudine. Stavudine (d4T) has a molecular weight ofered in the urine consists of the pharmacologically inac- 224.22 D, a fraction bound to plasma protein less thantive glucuronide metabolite (GZDV). Dosage adjust- 10%, and a volume of distribution of 0.5 to 1 L/kg. Fortyment is necessary for patients with renal insufficiency percent of a dose of d4T is excreted unchanged in the(Table 5). Hemodialysis has a minimal effect on ZDV urine by both glomerular filtration and tubular secretionelimination but enhances GZDV elimination significantly [21]. As there are alterations in the pharmacokinetics of(hemodialysis clearances of ZDV and GZDV using a d4T in patients with renal impairment, dosage adjust-

ment is necessary for these patients (Table 5). Hemodial-cuprophan membrane are 63 and 91 mL/min, respec-


Fig. 1. Molecular structure of antiretroviral drugs.

ysis clearance of d4T is relatively high (120 mL/min), dialysis patients, the daily dose should be administeredafter the dialysis session to avoid any drug loss. This iswith an extraction ratio of 68% during a four-hour hemo-

dialysis session on a nonspecified hemodialysis mem- consistent with the physicochemical and kinetic proper-ties of the drug (molecular weight �1000 D, and lowbrane (abstract; Fiedler et al, J Am Soc Nephrol 207:

1059A, 1998). Since FHD has not been evaluated in hemo- protein bound fraction and volume of distribution).


Table 1. Pharmacokinetic studies characteristics of antiretroviral drugs in hemodialyzed patients

Number Number ofDrug Ref. of studies patients S/M - dose - route Pharmacokinetics

Zidovudine [15, 40, 41, 42] 4 25 S - 200 mg - OralM - 100 mg tid - Oral ZDV: unchangedM - 100 mg � 4/day GZDV: T1⁄2

� 2 and AUC � 10M - 100 mg tid - Oral

Didanosine [17, 43] 2 24 S - 300 mg - Oral or IV T1⁄2� 3 and AUC � 7

Zalcitabine — — — —Lamivudine [(Fiedler),a 23, 24] 3 8 M - 150 mg bid - Oral

S - 300 mg - Oral T1⁄2increased and AUC � 5

M - 150 mg bid - OralStavudine (Fiedler)a 1 1 M - 30 mg bid - Oral T1⁄2

� 4 and AUC � 4Abacavir [26] 1 1 M - 300 mg daily - Oral UnchangedNevirapine [27] 1 1 M - 200 mg daily - Oral UnchangedDelavirdine — — — —Efavirenz [29] 1 1 M - 600 mg daily - Oral UnchangedRitonavir [32] 1 1 Rito.: M - 200 mg bid - Oral Rito.: UnchangedSaquinavir [32] Saqui.: M - 600 mg bid - Oral Saqui.: T1⁄2

� 2 and AUC � 5Indinavir [(Fiedler),a 34, 35] 3 3 M - 800 mg tid - Oral UnchangedNelfinavir — — — —Amprenavir — — — —

Abbreviations are: bid, twice daily; tid, thrice daily; S/M, single or multiple; HD, hemodialysis; T1⁄2, elimination half-life; AUC, area under the concentration-time

curve, ZDV, zidovudine; GZDV, zidovudine glucuromide; Rito, ritonavir; Saqui, saquinavir.a Fiedler et al, J Am Soc Nephrol 207:1059A, 1998

Table 2. Renal abnormalities induced by antiretroviral drugs

Drug Abnormality

Nucleoside reverse transcriptase inhibitorsZidovudine Lactic acidosis, rhabdomyolysisDidanosine Lactic acidosis, elevated serum uric acid, magnesium, reverse elevation of plasma albumin in

a hemodialyzed patientZalcitabine Acute renal failure, lactic acidosis, hyponatremia, hypocalcemia, gut, renal calculiStavudine Lactic acidosis, plasma dysuric acidLamivudine Lactic acidosisAbacavir —

Non-nucleoside reverse transcriptase inhibitorsNevirapine Lactic acidosisDelavirdine —Efavirenz —

Protease inhibitorsSaquinavir Lactic acidosis, hypocalcemia, dys-kaliemia, magnesemia and phosphoremia, pancreatorenal

syndromeRitonavir Acute renal failure, pancreatorenal syndrome, hypocalcemia, dys-kaliemia, phosphoremia,

magnesemia and uricemiaIndinavir Lactic acidosis, intratubular precipitation, urinary lithiasis, renal insufficiencyNelfinavir Lactic acidosis, hypocalcemia, dys-phosphoremia and uricemia, urinary lithiasisAmprenavir —

Lamivudine. Lamivudine (3TC) has a molecular weight the daily dose should be administered after the dialysissession to minimize drug loss. This is consistent withof 229.3 D, a fraction bound to plasma protein less than

36%, and a volume of distribution of 1.3 L/kg. Seventy the physicochemical and kinetic properties of the drug(molecular weight �1000 D, and low protein-bound frac-to 100% of each dose of 3TC is excreted unchanged in

the urine [22] by both glomerular filtration and tubular tion and volume of distribution).One of our HIV-infected and hemodialyzed patientssecretion, with a threefold to fourfold increase in the

drug’s half-life in patients with renal impairment [23]. was treated for six months with normal doses of lamivu-dine (150 mg every 12 hours), which resulted in excessiveTherefore, dosage adjustment is necessary for patients

with renal insufficiency (Table 5). Hemodialysis clear- drug accumulation but no side effects [25].Abacavir. Abacavir (Aba) has a molecular weight ofance of lamivudine is by no means negligible (106 mL/

min) on a polysulfone membrane. Its extraction ratio is 670.76 D (sulfate), a fraction bound to plasma proteinless than 15%, and a volume of distribution of 0.8 to 1.953% [24] and FHD is 25%. In hemodialyzed patients,


Table 3. Properties of nucleoside reverse transcriptase inhibitors affecting dialysis clearance

Properties

Molecular weight Protein binding Volume of distributionDrug Daltons % L/kg Primary elimination pathway

Zidovudine 267.24 7–38 1.4–1.6 Hepatic metabolismDidanosine 236.23 �5 0.7–0.9 Renal excretion/hepatic metabolismZalcitabine 211.22 �4 0.5–0.6 Renal excretionStavudine 224.22 �10 0.5–1 Renal excretion/hepatic metabolismLamivudine 229.3 �36 1.3 Renal excretionAbacavir 670.76 10–13 0.8–1.9 Hepatic metabolism

(sulfate)

Table 4. Pharmacokinetic parameters of nucleoside reverse transcriptase inhibitors in subjectswith normal renal function and hemodialysis (HD) patients

Area under theconcentration-time curve

Elimination half-life hours mg·h/L Hemodialysis session 4 hours% excreted

unchanged in Non-HD Non-HD Extraction ratio HD clearance FHD

Drug urine [45, 56] Normal day HD day Normal day HD day Membrane % mL/min %

Zidovudine 15–20 0.9–1.3 1.3–1.5 1.1 1.2–1.6 2.8–3.4 2.3–3.1 C 50–55 63 5–8Didanosine 50–55 1.6 3.1 3.1 5.7 14 1.8–3.2 C, PS, PAN 40–60 85–210 40Zalcitabine 70–75 2.2 10.7 — 18–34 93–189 — — 50 — —Stavudine 35–50 1.3 5–8 — 1.9 6 — — 68 120 —Lamivudine 70–100 5–9 13–20 13–17 13–18 60–104 60–100 PS 45–60 106 25Abacavir �20 1.2 2.08 — 6 7.91 — PS 24 60–80 10

Membranes are: C, cuprophane; PS, polysulfone; PAN, polyacrylonitrile. FHD is hemodialysis clearance on total body clearance.

L/kg. Less than 20% of Aba is excreted unchanged in 60% and a volume of distribution of 1.3 L/kg. Less than5% of the parent compound is excreted unchanged, andthe urine [26]. We studied the pharmacokinetics of Aba

in five HIV-infected patients who had various degrees 75 to 80% of each dose is recovered in the urine as theof renal failure with creatinine clearance ranging from glucuronide metabolite. We studied the pharmacokinet-60 to less than 10 mL/min [27]. No change was observed ics of nevirapine in two HIV-infected patients [28]. Onein the pharmacokinetics of Aba. Therefore, dosage ad- had a creatinine clearance of 60 mL/min and receivedjustment is not necessary for patients with renal insuffi- 200 mg of nevirapine twice daily. The other patient wasciency (Table 5). In our patients, hemodialysis clearance hemodialyzed and received a single 200 mg dose of nevir-of Aba and its extraction ratio on a polysulfone mem- apine. The pharmacokinetics of the drug were not sig-brane are 80 mL/min and 24%, respectively. FHD is 10%, nificantly modified by the renal insufficiency. However,and hemodialysis clearance is thus not clinically relevant the pharmacokinetics of the glucuronide metabolite havedespite physicochemical and kinetic properties, sug- not been studied. The hemodialysis clearance and extrac-gesting a good dialyzability (molecular weight �1000 D, tion ratio on a polysulfone membrane of nevirapine wereand low protein-bound fraction and volume of distribu- 100.5 mL/min and 46.5%, respectively. FHD was 88%.tion). Indeed, since hepatic metabolism of Aba is exten- This is consistent with the physicochemical and kineticsive, the fraction removed will not be clinically signifi- properties of the drug (molecular weight �1000 D, weakcant, and Aba may thus be administered regardless of protein-bound fraction and low volume of distribution).the hemodialysis schedule. These results suggest that no dosage adjustment is neces-

sary in renal insufficiency, but that nevirapine should beNon-nucleoside reverse transcriptase inhibitors administered after the dialysis session, to minimize drug

loss (Table 5).Except for nevirapine, protein binding of non-NRTIsis higher than normal, and the free fractions of these drugs Delavirdine. Delavirdine has a molecular weight of

456.57 D, and a fraction bound to plasma protein up toare lower, in renal insufficiency (Tables 6 and 7). Theyare primarily metabolized by the liver isoenzyme CYP3A4. 99%. Less than 5% of the drug is excreted unchanged

in the urine and about 51%, as the pharmacologicallyTheir excretion routes are both renal and biliary.Nevirapine. Nevirapine has a molecular weight of inactive metabolite. The disposition of delavirdine in

renal failure and hemodialysis has not been studied.266.3 D, a fraction bound to plasma protein less than


Table 5. Oral dosage recommendations for antiretroviral drugs in hemodialyzed patients

Drugs Normal dosage Hemodialyzed patients References

Nucleoside reverse transcriptase inhibitorsZidovudineb 200 mg q 8 h 100 mg q 8 h [18, 44–47]Didanosineb [18, 43, 44, 46]

BW �60 kg 200 mg q 12 h 100 mg q 24 hBW �60kg 125 mg q 12 h 50 mg q 24 h

Zalcitabineb 0.75 mg q 8 h 0.75 mg q 24 h [18, 43–46]Stavudineb [18, 43, 44, 46]

BW �60 kg 40 mg q 12h 20 mg q 24hBW �60 kg 30 mg q 12 h 15 mg q 24 h

Lamivudineb 150 mg q 12 h 150 mg � 1 [18, 43, 44, 46]then 25–50 mg q 24 h

Abacavirb 600 mg q 12 h Normal dosage [25]Non-nucleoside reverse transcriptase inhibitors

Nevirapineb 200 mg q 24 h 14 days Normal dosage [26]then 200 mg q 12 h

Delavirdineb 400 mg q 8 h NAEfavirenz 600 mg q 24 h Normal dosage [28]

Protease inhibitorsSaquinavir 600 mg q 8 h Normal dosagea [31, 45, 46]Ritonavir 600 mg q 12 h Normal dosage [31, 45, 46]Indinavir 800 mg q 8 h Normal dosage [33, 45, 46]Nelfinavir 750 mg q 8 h Normal dosage [37, 45, 46]Amprenavir 1200 mg q 12 h Normal dosage [37]

Abbreviations are: NA, not available; BW, body weighta When saquinavir is used in combination with ritonavir, its dose should be reducedb Drug should be administered after the hemodialysis session

Table 6. Properties of non-nucleoside reverse transcriptase schedule because of its extensive hepatic metabolisminhibitors affecting dialysis clearance

(Table 5). This is consistent with the physicochemicalProperties and kinetic properties of the drug (molecular weight

Molecular Protein Volume of �1000 D, and high protein-bound fraction and volumeweight binding distribution Primary elimination of distribution).

Drug Daltons % L/kg pathway

Nevirapine 266.3 50–60 1.2–1.4 Hepatic metabolism Protease inhibitorsDelavirdine 456.57 99 — Hepatic metabolism

Protease inhibitors (PIs) are mainly metabolized byEfavirenz 315.68 99–100 30 Hepatic metabolismthe liver and in the gastrointestinal tract by CYP iso-enzymes. Only a small percentage of these drugs is ex-creted by the kidney. One of their main characteristicsis their high protein binding, predominantly to �1-acidSince more than 95% of the drug is predominantly me-glycoprotein [31], except for indinavir. The pharmacoki-tabolized in the liver, no dosage adjustment should benetic profiles of PIs suggest that renal function shouldnecessary for patients with renal insufficiency (Table 5).not greatly affect the elimination of these drugs (TablesBecause hemodialysis removal of delavirdine has not8 and 9).been studied, the drug should be administered after the

Ritonavir. Ritonavir has a molecular weight of 720.95hemodialysis session to avoid any potential eliminationD, a fraction bound to plasma protein greater than 98%,by the dialysis.and a volume of distribution of 0.4 L/kg. Only 3 to 5%Efavirenz. Efavirenz has a molecular weight of 315.68of a dose of ritonavir is excreted unchanged in the urine.D, a fraction bound to plasma protein up to 99 to 100%,Since its renal clearance is negligible, a decrease in totaland a volume of distribution of 30 L/kg. Less than 1%body clearance is not likely to occur in patients withof the efavirenz dose is excreted unchanged in the urinerenal insufficiency. Ritonavir is highly bound to protein.[29]. The disposition of efavirenz in renal impairmentThus, hemodialysis is unlikely to remove a significanthas not been studied. However, no modification of itsamount of the drug [32]. We studied the pharmacokinet-pharmacokinetics should occur in these patients, becauseics of ritonavir in two HIV-infected patients [33]. Onemost of efavirenz is metabolized in the liver by CYP3A4.had a creatinine clearance of 20 mL/min, and the otherSince its hemodialysis clearance and extraction ratio onpatient was hemodialyzed. Both of them received 400a polysulfone membrane are low (20 mL/min and 16%,mg of ritonavir twice daily. The pharmacokinetics ofrespectively [30]) and despite that FHD is unknown, efavi-

renz may be administered regardless of the hemodialysis the drug were not significantly modified by the renal


Table 7. Pharmacokinetic parameters of non-nucleoside reverse transcriptase inhibitors in subjectswith normal renal function and hemodialysis (HD) patients

Area under theElimination half-life concentration-time curve

hours mg·h/L Hemodialysis session% excreted


Drug urine [45, 46] Normal day HD day Normal day HD day Membrane % mL/min %

Nevirapine �3–5 25–30 21.8 19.2 — 254.4 142 PS 46.5 100.5 88Delavirdine �5 4–11 — — — — — — — — —Efavirenz �1 40–55 — 55 58 — 40 PS 16 20 —

Table 8. Properties of protease inhibitors affecting with a FHD of 4%. This is consistent with the physico-dialysis clearance

chemical and kinetic properties of the drug (molecularProperties weight �1000 D, and high protein bound fraction and

Molecular Protein Volume of volume of distribution). These data suggest that no dos-weight binding distribution Primary elimination age adjustment is necessary in renal insufficiency when

Drug Daltons % L/kg pathwaysaquinavir is administered alone, but that the dosage

Ritonavir 720.95 �98 0.4 Hepatic metabolism reduction should be performed when used in combina-Saquinavir 766.95 �98 3.6–10 Hepatic metabolism

tion with ritonavir. The drug can be administered regard-Indinavir 711.88 60 14 Hepatic metabolismNelfinavir 663.9 �98 2–7 Hepatic metabolism less of the dialysis schedule (Table 5).Amprenavir 507.65 90 16.4 Hepatic metabolism Indinavir. Indinavir has a molecular weight of 711.88

D, a fraction bound to plasma protein of 60%, and avolume of distribution of 14 L/kg. Approximately 10%of a dose of indinavir is recovered unchanged in the

insufficiency. Thus, in hemodialyzed patients, the phar- urine [34]. In patients with renal insufficiency, the phar-macokinetic parameters of ritonavir are unchanged com- macokinetics of indinavir are not modified (abstract;pared to those in patients with normal renal function. Fiedler et al, J Am Soc Nephrol 207:1059A, 1998) [35].The hemodialysis clearance and extraction ratio of rito- In these studies, hemodialysis clearance and extractionnavir on a polysulfone membrane are 24 mL/min and ratio on a polysulfone membrane were very low, at 315.5%, respectively [32], with a FHD of 4%. This is consis- mL/min and 3%, respectively, much lower than thosetent with the physicochemical and kinetic properties of reported by Guardiola et al (185 7.2 mL/min) [36].the drug (molecular weight �1000 D, and high protein However, our results and those of Fiedler were not unex-bound fraction and volume of distribution). These data pected because they are consistent with the physicochemi-suggest that no dosage adjustment is required for patients cal and kinetic properties of the drug (molecular weightwith renal insufficiency and that the drug can be adminis- �1000 D, and high protein bound fraction and volumetered regardless of the dialysis schedule (Table 5). of distribution). Furthermore, FHD was almost 0% in

Saquinavir. Saquinavir has a molecular weight of our study. Therefore, indinavir can be administered to766.95 D, a fraction bound to plasma protein greater patients with renal insufficiency at a normal dosage, irre-than 98% and a volume of distribution of 3.6 and up to spective of their hemodialysis schedule (Table 5).10 L/kg. Only 1 to 3% of a dose of saquinavir is excreted Nelfinavir. Nelfinavir has a molecular weight of 663.9unchanged in the urine. Since renal clearance is negligi- D, a fraction bound to plasma protein greater than 98%,ble, a decrease in total body clearance is not likely to and a volume of distribution of 2 to 7 L/kg.occur in patients with renal insufficiency. Saquinavir is About 2% of nelfinavir is excreted unchanged in thehighly bound to protein. Thus, hemodialysis is unlikely urine, and its pharmacokinetics are not affected by renalto remove a significant amount of the drug. We studied insufficiency. Furthermore, there are two observationsthe pharmacokinetics of saquinavir in two HIV-infected of hemodialyzed patients who have been treated withpatients. One had a creatinine clearance of 20 mL/min, normal doses of nelfinavir (1500 to 2250 mg/day) forand the other patient was hemodialyzed. Both of them three months and have not experienced any side-effectsreceived 600 mg of saquinavir twice daily. The pharmaco- [37] (abstract; Bhatti et al, 12th World AIDS Conference,kinetics of the drug were significantly modified, but this Geneva, Switzerland 28 June through 3 July 1998, ab-was related to the combination with ritonavir more than stract 60541). Therefore, dosage adjustment is not neces-the alteration of the renal function. The hemodialysis sary for such patients (Table 5) [38]. While hemodialysisclearance and extraction ratio of saquinavir on a polysul- clearance of nelfinavir has not been evaluated, the physi-

cochemical and kinetic data suggest that nelfinavir isfone membrane are 3.25 mL/min and 2%, respectively,


Table 9. Pharmacokinetic parameters of protease inhibitors in subjects with normal renal function and hemodialysis (HD) patients

Area under theElimination half-life concentration-time curve

hours mg·h/L Hemodialysis session% excreted


Drug urine [45, 46] Normal day HD-day Normal day HD day Membrane % mL/min %

Saquinavir 1–3 7–13 12.72 9.72 — 109.16a 93.24a PS 2 3.25 4Ritonavir �5 3–5 5.8 5.13 — 81.26a 60.72a PS 15.5 24 4Indinavir 10–20 1.4–2.2 2.1 1.03 14–28 26.6 18.2 PS 3 3.25 almost 0Nelfinavir 1–2 3.5–5 4 4 30 — — — — — —Amprenavir �20 9 — — — — — — — — —

a Values for combination Ritonavir-Saquinavir treatment

Table 10. Side effects of antiretroviral therapies [44]

Drug Side effects

Nucleoside reverse transcriptase inhibitorsZidovudine Gastrointestinal disorders, anemia, neutropenia, myopathyDidanosine Diarrhea, nausea, pancreatitis, neuropathyZalcitabine Neuropathy, oral ulcerationsStavudine NeuropathyLamivudine Well toleratedAbacavir Hypersensitivity: forbids any re-introduction of the treatment

Non-nucleoside reverse transcriptase inhibitorsNevirapine Rash, hepatitisDelavirdine Headache, gastrointestinal disorders, rashEfavirenz Dizziness, headache, insomnia, rash

Protease inhibitorsSaquinavir NauseaRitonavir Gastrointestinal disorders, hepatitis, hypertriglyceridemia, perioral paresthesiaIndinavir Urinary lithiasis, gastrointestinal disorders, hyperbilirubinemiaNelfinavir Gastrointestinal disorders (diarrhea), rashAmprenavir Gastrointestinal disorders, rash, headache, paresthesia

weakly dialyzable. Indeed, its molecular weight is less Second, simple and easy-to-apply recommendationsthan 1000 D, but nelfinavir is highly bound to plasma should be established for the treatment of patients un-protein and has a large volume of distribution (greater dergoing hemodialysis. These recommendations cannotthan 2 L/kg). be determined without a pharmacokinetic approach to

Amprenavir. Amprenavir has a molecular weight of the drug elimination during hemodialysis. We emphasize507.65 D, a fraction bound to plasma protein of 90%, that except for ZVC and dianosine, more primary dataand a volume of distribution of 16.4 L/kg. A recent PI, are required on the pharmacokinetics of antiretroviralits metabolism is primarily hepatic, and its renal elimina- drugs in patients with renal insufficiency.tion is minimal. We have shown that the pharmacokinetics

Reprint requests to Hassane Izzedine, M.D., Department of Nephrol-of amprenavir do not change in patients with renal insuf-ogy, Pitie Salpetriere Hospital, 47-83 Boulevard de l’Hopital, 75013 Paris,

ficiency (creatinine clearance of 20 mL/min) [39]. There- France.E-mail: [email protected], no dosage adjustment is necessary for those patients

(Table 5). Amprenavir clearance by hemodialysis hasnot been evaluated. However, the physicochemical and REFERENCESkinetic data suggest that amprenavir is weakly dialyzable. 1. Kimmel PL, Phillips TO, Furreira-Centeno A, et al: HIV-associ-Indeed, molecular weight is less than 1000 D, and am- ated immune-mediated renal disease. Kidney Int 44:1327–1340, 1993

2. Casanova S, Mazzucco G, di Belgiojoso B, et al: Pattern ofprenavir is highly bound to plasma protein and has aglomerular involvement in human immunodeficiency virus-infectedlarge volume of distribution (greater than 2 L/kg). patients: An Italian study. Am J Kidney Dis 26:446–453, 1995

3. Rowland M, Tozer TN: Dialysis, in Clinical Pharmacokinetics:Concepts and Applications (3rd ed), Baltimore, Lippincott, Wil-

CONCLUSION liams, and Wilkins, 1995, pp 443–4624. Reetze-Bonorden P, Bohler J, Keller E: Drug dosage in patientsIn patients with renal insufficiency, careful dosage ad-

during continuous renal replacement therapy. Clin Pharmacokinetjustment is mandatory to optimize the exposure to the 24:362–379, 1993

5. Reidenberg MM: The binding of drugs to plasma proteins anddrug and reduce the risk of adverse effects (Table 10).


the interpretation of measurements of plasma concentrations of 25. Izzedine H, Launay-Vacher V, Deray G: Dosage of lamivudinein a haemodialysis patient. Nephron 86:553, 2000drugs in patients with poor renal function. Am J Med 62:466–470,

1977 26. Fister RH, Faulds D: Abacavir. Drugs 55:729–736, 199827. Izzedine H, Launay-Vacher V, Aymard G, et al: Pharmacokinet-6. Mc Namara PJ, Lalka D, Gibaldi M: Endogenous accumulation

products and serum protein binding in uremia. J Lab Clin Med ics of abacavir in HIV1-infected patients with impaired renal func-tion. Nephron 87:186–187, 200198:730–740, 1981

7. Boobis SW: Alteration of plasma albumin in relation to decreased 28. Izzedine H, Launay-Vacher V, Aymard G, et al: Pharmacokinet-ics of nevirapine in haemodialysis. Nephrol Dial Transplant 16:192–drug binding in uremia. Clin Pharmacol Ther 22:147–153, 1977

8. Rowland M, Tozer TN: Basic considerations, in Clinical Pharma- 193, 200129. Efavirenz package insert. Paris, Dupont Pharma Pharmaceuticals,cokinetics: Concepts and Applications (3rd ed), Baltimore, Lippin-

cott, Williams, and Wilkins, 1995, pp 11–17 199830. Izzedine H, Aymard G, Hamani A, et al: Pharmacokinetics of9. Reidenberg MM: The biotransformation of drugs in renal failure.

Am J Med 62:482–485, 1977 efavirenz in a patient on maintenance hemodialysis. AIDS 14:618–619, 200010. Keller F, Wilms H, Schultz G, et al: Effect of plasma protein

binding, volume of distribution and molecular weight on the frac- 31. Barry M, Gibbons S, Back D, Mulcahy P: Protease inhibitors inpatients with HIV disease: Clinically important pharmacokinetiction of drugs eliminated by hemodialysis. Clin Nephrol 19:201–205,

1983 considerations. Clin Pharmacokinet 30:194–209, 199732. Lee AP, Faulds D: Ritonavir. Drugs 52:541–546, 199611. Maher JF: Pharmacokinetics in patients with renal failure. Clin

Nephrol 21:39–46, 1984 33. Izzedine H, Launay-Vacher V, Legrand M, et al: Pharmacokinet-ics of ritonavir and saquinavir in haemodialysis. Nephron 87:186–12. Golper TA Marx MA, Shuler C, Benett WM: Drug dosage in

dialysis patients, in Replacement of Renal Function by Dialysis (4th 187, 200134. Baclani SK, Woolf EJ, Hoagland VL, et al: Disposition of indi-ed), edited by Jacobs C, Kjellstrand CM, Koch CM, Winchester

JF, London, Kluwer Academic Publishers, 1996, pp 750–820 navir, a potent HIV-1 protease inhibitor, after an oral dose inhumans. Drug Metab Dispos 24:1389–1394, 199613. Rutstein DD, Castelli WP, Nickerson R: Heparin and human

lipid metabolism. Lancet 1:1003, 1969 35. Izzedine H, Aymard G, Hamani A, et al: Indinavir pharmacokinet-ics in haemodialysis patients. Nephrol Dial Transplant 15: 1102–14. Lee CS, Marbury TC: Drug therapy in patients undergoing haemo-

dialysis: Clinical pharmacokinetical considerations. Clin Pharma- 1103, 200036. Guardiola JM, Magues MA, Domingo P, et al: Indinavir pharma-cokinet 9:42–66, 1984

15. Sommadossi JP: Nucleoside analogs: Similarities and differences. cokinetics in haemodialysis dependent end stage renal failure.AIDS 12:1395, 1998Clin Infect Dis (Suppl 1)56:S7–S15, 1993

16. Singlas E, Pioger JC, Taburet AM, et al: Zidovudine disposition 37. Wali RK, Drachenberg CI, Papadimitriou JC, et al: HIV-1 associ-ated nephropathy and response to highly-active antiretroviral ther-in patients with renal impairment: Influence of hemodialysis. Clin

Pharmacol Ther 46:190–197, 1989 apy. Lancet 352:783–784, 199838. Izzedine H, Diquet B, Launay-Vacher V, et al: Pharmacokinetics17. Garatto R, Casanto-Viguier E, Barillon J, et al: Influence of

hemodialysis on zidovudine and its glucuronide (GAZT) pharma- of nelfinavir in an HIV patient with renal insufficiency. AIDS13:1989, 1999cokinetics: Two case reports. Int J Clin Pharmacol Ther Toxicol

27:535–539, 1989 39. Izzedine H, Launay-Vacher V, Legrand M, et al: Pharmacokinet-ics of amprenavir in HIV1-infected patients with renal dysfunction.18. Singlas E, Taburet A, Lebas B, et al: Didanosine pharmacokinet-

ics in patients with normal and impaired renal function: Influence Nephrol Dial Transplant (in press)40. Pachon J, Cisneros JM, Castillo JR, et al: Pharmacokinetics ofof hemodialysis. Antimicrob Agents Chemother 36: 1519–1524, 1992

19. Bazunga M, Tran HT, Kertland H, et al: The effects of renal zidovudine in end stage renal disease: Influence of hemodialysis.AIDS 6:827–830, 1992impairment on the pharmacokinetics of zalcitabine. J Clin Pharma-

col 38:28–33, 1998 41. Tartaglione TA, Holeman E, Opheim K, et al: Zidovudine disposi-tion during hemodialysis in a patient with acquired immunodefi-20. Hilts AE, Fish DN: Dosage adjustment of antiretroviral agents

in patients with organ dysfunction. Am Health Syst Pharm 55:2528– ciency syndrome. J Acquir Immune Defic Syndr 3:32–34, 199042. Deray G, Diquet B, Martinez F, et al: Pharmacokinetics of zido-2533, 1998

21. Creton EM, Zhon Z, Kidd LB, et al: In vitro and in vivo disposition vudine in a patient on maintenance hemodialysis. N Engl J Med319:1606–1607, 1988and metabolism of 3-deoxy-2’,3-didehydrothymidine. Antimicrob

Agents Chemother 37:1916–1925, 1993 43. Knupp CA, Hak LJ, Coakley DF, et al: Disposition of didanosinein HIV-seropositive patients with normal renal function or chronic22. Van Leeuwen R Lange JMA, Hussey EL, et al: The safety and

pharmacokinetics of a reverse transcriptase inhibitor, 3TC, in pa- renal failure: Influence of hemodialysis and continuous ambulatoryperitoneal dialysis. Clin Pharmacol Ther 60:535–542, 1996tients with HIV infection: A phase I study. AIDS 6:1471–1475,

1992 44. Jayasekara D Aweeka FT, Rodriguez R, et al: Antiviral therapyfor HIV patients with renal insufficiency. AIDS 21:384–395, 199923. Perzy CM, Farelds D: Lamivudine: A review of its antiviral activ-

ity, pharmacokinetic properties and therapeutic efficacy in the 45. Raffi F: Anti-HIV drugs. La Revue du Praticien 49:763–771, 199946. Gurtman A, Borrego F, Klotman ME: Management of antiret-management of HIV infection. Drugs 59:550–558, 1997

24. Johnson MA, Vertooten GA, Daniel MJ, et al: Single dose phar- roviral therapy. Semin Nephrol 18:459–480, 199847. Sreepada Rao TK: Acute renal failure syndromes in human immu-macokinetics of lamivudine in subjects with impaired renal function

and the effect of hemodialysis. Br J Clin Pharmacol 46:21–27, 1998 nodeficiency virus infection. Semin Nephrol 18:378–395, 1998

an appraisal of antiretroviral drugs in hemodialysis

Documents